Scholarlink

<p>Thyroid-stimulating hormone receptor antibody plays a crucial role in the diagnosis and management of Graves’ disease. Artificial intelligence has demonstrated the ability to identify patterns in various medical fields. This exploratory study aimed to investigate the feasibility of applying artificial intelligence to analyze the levels of thyroid-stimulating hormone receptor antibody in patients with Graves’ disease. This case series analyzed the medical records of 50 patients with Graves’ disease who underwent total thyroidectomy between February 2006 and July 2022. Clinical data, including comorbidities, family history, and laboratory measurements, were collected to assess prognosis. A convolutional neural network model was used to analyze time-series antibody data. Python programming was employed to implement and interpret the models, and the results were rigorously validated against conventional statistical software. Patients were followed for an average of 6.9&#xa0;years. Among the 50 patients, 16 (32%) were male, and 5 (10%) had a family history of thyroid dysfunction. The mean initial antibody level was 37.1&#xa0;IU/L, well above the normal range. A convolutional neural network identified an exponential decay model describing antibody dynamics over time. Subgroup analysis revealed a significant association between initial antibody levels and the presence of thyroid eye disease (<i>p</i> = 0.039). A Random Forest classifier predicted eye disease with high performance (accuracy: 0.90; area under the curve: 0.88; precision: 0.75; recall: 0.65; F1-score: 0.70). No recurrence was observed during follow-up. Artificial intelligence-assisted modeling of thyroid-stimulating hormone receptor antibodies may provide a novel approach to predicting clinical outcomes and guiding care in Graves’ disease.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Artificial Intelligence for Analyzing the Dynamics of TSH Receptor Antibodies in Graves’ Disease

  • 期刊论文

  • 发表日期：

    2025年11月25日

查看全文
  • Yong Joon Suh

摘要

Thyroid-stimulating hormone receptor antibody plays a crucial role in the diagnosis and management of Graves’ disease. Artificial intelligence has demonstrated the ability to identify patterns in various medical fields. This exploratory study aimed to investigate the feasibility of applying artificial intelligence to analyze the levels of thyroid-stimulating hormone receptor antibody in patients with Graves’ disease. This case series analyzed the medical records of 50 patients with Graves’ disease who underwent total thyroidectomy between February 2006 and July 2022. Clinical data, including comorbidities, family history, and laboratory measurements, were collected to assess prognosis. A convolutional neural network model was used to analyze time-series antibody data. Python programming was employed to implement and interpret the models, and the results were rigorously validated against conventional statistical software. Patients were followed for an average of 6.9 years. Among the 50 patients, 16 (32%) were male, and 5 (10%) had a family history of thyroid dysfunction. The mean initial antibody level was 37.1 IU/L, well above the normal range. A convolutional neural network identified an exponential decay model describing antibody dynamics over time. Subgroup analysis revealed a significant association between initial antibody levels and the presence of thyroid eye disease (p = 0.039). A Random Forest classifier predicted eye disease with high performance (accuracy: 0.90; area under the curve: 0.88; precision: 0.75; recall: 0.65; F1-score: 0.70). No recurrence was observed during follow-up. Artificial intelligence-assisted modeling of thyroid-stimulating hormone receptor antibodies may provide a novel approach to predicting clinical outcomes and guiding care in Graves’ disease.